GRBM lock in vce_v3_0_check_soft_reset(void *handle)

[StDenis, Tom]

Ok I'll submit a patch in a bit.


Tom


________________________________
From: Deucher, Alexander
Sent: Thursday, August 11, 2016 09:53
To: StDenis, Tom; amd-gfx list
Subject: RE: GRBM lock in vce_v3_0_check_soft_reset(void *handle)


Yes, it sounds like a bug.  It should take the mutex locally.



From: amd-gfx [mailto:amd-gfx-bounces at lists.freedesktop.org] On Behalf Of StDenis, Tom
Sent: Thursday, August 11, 2016 9:42 AM
To: amd-gfx list
Subject: GRBM lock in vce_v3_0_check_soft_reset(void *handle)



I'm working on tidying up vce v3 and came across the function



static int vce_v3_0_check_soft_reset(void *handle)



Which writes to the GRBM_GFX_INDEX but doesn't take the mutex around it.  Nor does the parent function in amdgpu_device.c which calls all of the check_soft_reset callbacks.



Is this a bug?



Tom
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/amd-gfx/attachments/20160811/f6cff91d/attachment.html>